RICE prioritization: the formula, a worked example and the mistakes that break it

RICE prioritization is a scoring method that ranks ideas by Reach × Impact × Confidence ÷ Effort: how many people an idea affects in a given period, how much it changes things for each of them, how sure you are about those two estimates and how much work it takes. The result is a rough "value per unit of effort" number that puts very different ideas on one scale. More importantly, it makes the reasoning behind a priority visible, so people argue about the estimates instead of about whose idea it is. The method was popularized by Intercom's product team. Treat it as a tool for a better conversation, not a machine that makes the decision for you.

The RICE formula

• Reach: how many people or accounts the idea affects within a fixed period, such as a quarter. Take it from analytics where you can.
• Impact: how much the idea moves your goal for each person it reaches.
• Confidence: how much evidence stands behind the reach and impact estimates, as a percentage.
• Effort: the total work from everyone involved (product, design, engineering, marketing, support), in person-months.

The first three multiply value; effort divides it. Double the effort and the score halves; halve the confidence and it halves too.

What the scale values mean

Reach and effort are counts. Impact and confidence use fixed steps, so nobody can quietly score an idea "2.7":

Three rules keep the numbers comparable. Use the same period for every item's reach. Score impact against one goal, the objective on your roadmap, not "value" in general. And agree in advance what "high" impact means, with an example from past work. An idea below 50% confidence is a bet, not a plan: run a cheap experiment first.

A worked example: six ideas, one quarter

A Ukrainian B2B SaaS helps about 4,000 small online shops manage orders and deliveries. This year's objective is to keep more shops paying, measured by the share of new shops still paying after 90 days. The team has six candidates for next quarter. Reach is shops per quarter; impact is how much an idea moves a shop it reaches towards staying.

Scores 4 and 6 are rounded (266.67 and 93.75). Here is one row in full, bulk labels:

Where the estimates came from:

• Onboarding checklist: the team's data shows that shops which finish setup in their first week stay far more often, so confidence is 80%. It's small: one person-month.
• Bulk labels: shops that ship dozens of orders a day ask for it in support tickets every week. It saves them time daily, but on its own it rarely decides whether a shop stays: impact 1.
• Abandoned-cart messages: could matter a lot to shops that sell through their own storefront (impact 2), but nobody has checked, so confidence is 50%.
• Dashboard redesign: every shop sees it, but hardly anyone stays or leaves because of a chart: impact 0.25.
• Marketplace stock sync: high impact for the 900 shops that sell on marketplaces, but big, and uncertain.
• Mobile app: the loudest request, with low impact on staying and the largest effort.

Reading the result

The top two are clear: small, well understood and backed by evidence. Below them, the gaps are smaller than they look. Every score multiplies four estimates, so a difference of 400 against 267 says less than the digits suggest. A useful habit: treat scores that are close as ties and let other arguments decide.

Then ask which single estimate would change the order. If the onboarding checklist takes 3 person-months instead of 1, its score drops from 1,200 to 400, level with abandoned-cart messages. That effort estimate is the one to check before committing. If a quick prototype raised confidence in marketplace sync to 80%, its score would go from 150 to 240: still below the dashboard's 267, so, practically, a tie. A cheap experiment that raises confidence is often worth more than another meeting about the scores.

The mobile app's score also does a job: it turns a loud request into an explainable "not now". Saying no is easier when the reasons are on the table.

Pitfalls that break RICE prioritization

False precision

A score of 266.67 looks exact, but it's the product of four guesses. Round the scores, group them into bands (high, medium, low) and never argue about decimals.

Gaming the scores

Whoever wants an idea tends to rate its impact a 3 and its confidence 100%. Guard against it: define the scale with real examples, score as a group rather than alone, ask for evidence behind any confidence above 50%, and compare estimates with actual results after launch, so the team's guesses get better over time.

Comparing items of different kinds

RICE ranks items that serve the same goal. Put a retention feature, a legal requirement, a database upgrade and a bet on a new market on one scale and the ranking means nothing: a law you must follow isn't "impact 2". Sort work into buckets first. Must-dos (legal, security, enabling work) go ahead by definition; split the remaining capacity between buckets; use RICE inside each one. Watch for reach bias too: RICE favors ideas that touch many people over deep value for a small but important segment, such as your largest customers.

Ignoring urgency

RICE has no time dimension. A seasonal window or a contract deadline is invisible to it, which is why cost of delay exists (see below).

RICE on an outcome-based roadmap

On an outcome-based roadmap, RICE works at two levels. Inside a theme, it helps choose which candidate solution to try first; in the example, all six ideas serve one objective. Across themes, it can order the queue in the Next column of a Now Next Later roadmap, as long as every theme is scored against the same objective. In both cases, estimate impact against the theme's signal, and the scores stay tied to a result instead of to opinions.

Alternatives worth knowing

• Value vs effort: rate each item's value and effort, then plot them on a two-by-two grid or divide one by the other. Fewer numbers, quicker to run, good for a first pass.
• Cost of delay: how much value you lose for each week an item isn't done, for example revenue missed before a seasonal peak. It makes urgency explicit. WSJF (weighted shortest job first) divides cost of delay by job size and takes the highest ratio first.
• MoSCoW: sorts a release's scope into Must have, Should have, Could have and Won't have (this time). It communicates priorities once they're decided; it doesn't decide them.
• ICE: Impact, Confidence and Ease, a lighter cousin of RICE without reach, popular with growth teams running many small experiments.

Key takeaways

• RICE prioritization scores each idea as Reach × Impact × Confidence ÷ Effort: value per unit of work.
• Use fixed scales for impact and confidence, one period for reach and one goal for impact.
• Scores are rough: round them, treat close scores as ties and find the estimate that would change the order.
• Don't compare items of different kinds on one scale; handle must-dos and urgency separately.
• Value vs effort, cost of delay and MoSCoW answer different questions and work well alongside RICE.

FAQ

What does RICE stand for?

Reach, Impact, Confidence and Effort. The first three estimate the value of an idea and how sure you are about it; effort is the cost. The score is the first three multiplied together and divided by effort.

What is a good RICE score?

There is no absolute benchmark. A score only means something compared with other ideas scored by the same team, with the same scales, the same reach period and the same goal. A 960 in one team and a 960 in another aren't comparable.

Should I just build the idea with the highest RICE score?

Not automatically. The ranking is a starting point for a discussion: check which estimates are shakiest, what would change the order and whether something urgent or mandatory sits outside the scoring. Then decide, and write down why.

Learn it hands-on

Start with the free lesson What a roadmap is (and isn't): a CEO asks for every feature with a month next to it, and you practice answering with direction instead of dates. Module 4 of the free course Outcome-Based Product Roadmaps is about prioritizing: Why gut feeling fails covers criteria and bias, Scoring frameworks walks through value vs effort and RICE, and Trade-offs and saying no shows how to turn a ranking into a decision people accept.